HAL-CentraleSupelec
Not a member yet
77624 research outputs found
Sort by
Deep learning with fourier features for regressive flow field reconstruction from sparse sensor measurements
No full textInternational audienceMany applications in computational and experimental fluid mechanics require effective methods for reconstructing the flow fields from limited sensor data. However, this task remains a significant challenge because the measurement operator that provides the punctual sensor measurement for a given state of the flow field is often ill-conditioned and non-invertible. This issue impedes the feasibility of identifying the forward map, which is, theoretically, the inverse of the measurement operator, for field reconstruction purposes. While data-driven methods are available, their generalizability across different flow conditions (e.g., different Reynold numbers) remains questioned. Moreover, they frequently face the problem of spectral bias, which leads to smooth and blurry reconstructed fields, thereby decreasing the accuracy of reconstruction. We introduce FLRNet, a deep learning method for flow field reconstruction from sparse sensor measurements. FLRNet employs a variational autoencoder with Fourier feature layers and incorporates an extra perceptual loss term during training to learn a rich, low-dimensional latent representation of the flow field. The learned latent representation is then correlated to the sensor measurement using an attention-based network. We validated the reconstruction capability and the generalizability of FLRNet under various fluid flow conditions and sensor configurations, including different sensor counts and sensor layouts. Numerical experiments show that in all tested scenarios, FLRNet consistently outperformed other baselines, delivering the most accurate reconstructed flow field and being the most robust to noise
New Structural Insights and High Temperature Phase Transitions in Bismuth and Lead Uranates
No full textInternational audienceDespite extensive implementation of uranium and its oxides throughout the nuclear fuel cycle, studies of the ternary uranate structures are limited. The Bi-U-O and Pb-U-O systems are at the forefront of recent scientific development in the context of the lead bismuth eutectic cooled fast reactors. Given the social importance and the gap in our understanding of the solid state and crystal chemistry of uranium oxides, and the development of modern robust crystallographic techniques, there is compelling reasoning to return to studying uranium oxides.Bi2UO6 is one of few known structures in the Bi-U-O system. Our investigation of the thermal response of Bi2UO6 has demonstrated the inclusion of superlattice reflections not previously reported [1], which we have linked to the formation of oxygen vacancies. Using high resolution synchrotron X-ray powder diffraction and Neutron Powder Diffraction (NPD) in combination with DFT studies from first principles, we have resolved a new cell and accurate structure for Bi2UO6, crystallising in space group P-3m1. PbUO4 is a well-known member of the Pb-U-O system with orthorhombic Pbcm symmetry as determined by single crystal X-ray diffraction.[2] Previous studies of the thermal response of PbUO4 conditions have revealed little structural changes [3], however our investigations have demonstrated the irreversible transformation of PbUO4 to an unreported fluorite-related phase, nominally ‘Pb2U2O7’. Using a combination of NPD and DFT calculations, we have resolved the structure of Pb2U2O7 as the orthorhombic pyrochlore in space group Imma.These findings contribute toward bridging the gap of understanding regarding the structures and response to external stimuli of some of the simplest uranium oxides. Establishing such knowledge of fundamental uranium chemistry has implications for both the nuclear and crystallographic fields. With further adoption of nuclear energy, there is a growing need for a better understanding in uranium chemistry to facilitate the safe handling and storage of spent nuclear fuels and potential alteration phases. As it follows, establishing a strong understanding of uranium crystal chemistry, exhibiting 5f electron chemistry, may provide unique insights relevant to functional material design, as is demonstrated by the interesting capabilities of thermally induced oxygen defect ordering not previously realised.[1] Koster, A.; Renaud, J.; Rieck, G. The crystal structures at 20 and 1000° C of bismuth uranate, Bi2UO6. Structural Science 1975, 31 (1), 127-131.[2] Cremers, T.; Eller, P.; Larson, E.; Rosenzweig, A. Single-crystal structure of lead uranate (VI). Crystal Structure Communications 1986, 42 (12), 1684-1685.[3] Popa, K.; Beneš, O.; Staicu, D.; Griveau, J.-C.; Colineau, E.; Raison, P.; Vigier, J.-F.; Pagliosa, G.; Sierig, M.; Vălu, O. Thermal properties of PbUO4 and Pb3UO6. Journal of Nuclear Materials 2016, 479, 189-19
Complexité paramétrée de relations entre les sous-shifts multidimensionnels
No full textWe study the parametrized complexity of fundamental relations between multidimensional subshifts, such as equality, conjugacy, inclusion, and embedding, for subshifts of finite type (SFTs) and effective subshifts. We build on previous work of E. Jeandel and P. Vanier on the complexity of these relations as two-input problems, by fixing one subshift as parameter and taking the other subshift as input. We study the impact of various dynamical properties related to periodicity, minimality, finite type, etc. on the computational properties of the parameter subshift, which reveals interesting differences and asymmetries.Among other notable results, we find choices of parameter that reach the maximum difficulty for each problem; we find nontrivial decidable problems for multidimensional SFT, where most properties are undecidable; and we find connections with recent work relating having computable language and being minimal for some property, showing in particular that this property may not always be chosen conjugacy-invariant.</p
A Comprehensive Assessment of Cancer Patient Performance Status Documentation in a Large, Multicentre Hospital System
No full textInternational audienceBackground: The performance status (PS) is an indicator of a cancer patient's ability to perform everyday activities and plays a key role in oncology. Research suggests that the documentation of PS scores in electronic health records (EHR) is deficient. Methods: We analysed PS score documentation (Karnofsky or ECOG/Zubrod/WHO) in the hospital, consultation, and multidisciplinary team meeting (MDT) records of patients newly referred for a cancer at a large, public, multisite hospital system, between 1 January 2019 and 1 June 2021. We developed a regular expression (RegEx) to automatically identify PS in documents and assessed what patient and hospital characteristics were associated with PS documentation. Results: Our RegEx achieved accuracy, and weighted-and macro-average F1 score, > 0.95 for all document types. We included 68,479 patients. 35% had a documented PS between -90 and +365 days of their first ICD-10 cancer code. 18% of MDT reports contained a PS score. In multivariate analysis, without accounting for metastatic status at diagnosis, odds ratios (ORs) for PS documentation in patient files varied by cancer type, from 0.47 (95% confidence interval: [0.42; 0.52]) for genitourinary to 3.30 [3.00; 3.61] for lung cancer, and hospital, from 0.27 [0.23; 0.33] to 3.38 [3.14; 3.63]. Male patients were more likely to have a documented PS (OR = 1.08 [1.04; 1.13]), as well as older patients. The number of each type of document was positively correlated with the presence of a score. When adding metastatic status at diagnosis, the OR for metastatic status was large (3.29 [3.13; 3.46]), but associations with other covariates were not noticeably affected. Documented PS close to diagnosis was associated with poorer 1-year survival (25% of patients with PS died within 1 year, vs 12% without PS). Conclusion: PS score documentation was variable and generally low. Improved documentation is required if EHRs are to be used as a source of real-world data
RAM-VQA: Restoration Assisted Multi-Modality Video Quality Assessment
No full textInternational audienceVideo Quality Assessment (VQA) strives to computationally emulate human perceptual judgments and has garnered significant attention given its widespread applicability. However, existing methodologies face two primary impediments:(1) limited proficiency in evaluating samples at quality extremes (e.g., severely degraded or near-perfect videos), and (2) insufficient sensitivity to nuanced quality variations arising from a misalignment with human perceptual mechanisms. Although vision-language models offer promising semantic understanding, their reliance on visual encoders pre-trained for high-level tasks often compromises their sensitivity to low-level distortions. To surmount these challenges, we propose the Restoration-Assisted Multi-modality VQA (RAM-VQA) framework. Uniquely, our approach leverages video restoration as a proxy to explicitly model distortion-sensitive features. The framework operates through two synergistic stages: a prompt learning stage that constructs a quality-aware textual space using triple-level references (degraded, restored, and pristine) derived from the restoration process, and a dual-branch evaluation stage that integrates semantic cues with technical quality indicators via spatio-temporal differential analysis. Extensive experiments demonstrate that RAM-VQA achieves state-of-the-art performance across diverse benchmarks, exhibiting superior capability in handling extremequality content while ensuring robust generalization.</div
Medispeech: a French Reading and Spontaneous Speech Corpus for Sleepiness Estimation
No full textInternational audienceExcessive Daytime Sleepiness (EDS) is associated with several diseases and therefore negatively affects the daily life of impacted people. Its diagnosis and follow-up are difficult because they require testing at the hospital for one full day. Monitoring patients regularly in ecological conditions may be done through speech analysis. Although several corpora containing speech from sleepy subjects exist, they do not suit ecological requirements regarding either the device used for recording or the speech elicitation tasks. In this paper, we introduce the Medispeech corpus containing reading, daily-life semi-spontaneous, and medically-oriented spontaneous tasks. Fifty-nine French subjects were recorded with both a professional-quality microphone and a smartphone using a dedicated application, resulting in 1,729 recordings for a total duration of 21 hours. Their EDS diagnosis was assessed by both a physiological objective measurement (mean sleep latency measured during a clinical test) and a subjective questionnaire (Karolinska Sleepiness Scale). Phenotyping of subjects is assured by collecting socio-demographic and medical data related to diverse dimensions of sleepiness, comorbidities, and addictions. Finally, we analyse the validity of our data collection protocol by measuring the effective duration of speech (after discarding pauses) and assessing its links with the collected subjects' characteristics
Anchor-based Haptics: Portable Interfaces Leveraging Real Surfaces to Display Grounded Forces for 3D Shape Simulation in VR
No full textInternational audienceHaptic interfaces enable shape simulation in virtual reality, making such content "tangible". Among existing haptic technologies, "grounded" interfaces simulate convincing grounded surfaces (like walls, tables, or other grounded shapes) but have limited workspaces, while "portable" interfaces are usable everywhere but struggle to display force feedback. In this paper, we formalize an alternative approach that we call "anchor-based haptics", which consists of portable interfaces capable of temporarily making contact with a real surface and leveraging it for grounded force feedback. To illustrate this approach, we present two new interfaces: T-Hank, which uses a reel-based extension mechanism to give height feedback from the surface, and R-Hank, which provides orientation feedback. The design of these interfaces is presented, and their potential interest in virtual reality has been investigated within two user studies by comparing them with vibrations in various 3D tasks. Overall, T-Hank and R-Hank enhance user experience for the exploration of 3D shapes and softness sensations. Taken together, results promote the use of the anchor-based haptic approach for VR applications in which both a large workspace and shape simulation play a prominent role
Electrocaloric cooling efficiency: Comparative insights on P(VDF-TrFE-CFE) polymer and BSTM ceramics
No full textInternational audienceIn response to the increasing demand for efficient and compact refrigeration and energy conversion devices, research has focused on identifying optimal electrocaloric (EC) materials among ferroelectric ceramics and polymers. This study investigates the EC properties of the Poly [(Vinylidene Fluoride) 0.664 -(Trifluoroethylene) 0.245 -(Chlorofluoroethylene) 0.091 ] terpolymer and multi-layer Ba 0.6 Sr 0.4 Ti 0.998 Mn 0.002 O 3 (BSTM) ceramics, comparing various parameters to assess their suitability for advanced energy applications.The multilayer ceramic capacitor contains a large amount of inactive material, which hinders the performance of the capacitor both in terms of ΔT ad and efficiency. Finite-element modeling with direct temperature measurement was therefore employed to extract intrinsic electrocaloric response from geometric and diffusion effects, providing the ΔT ad and diffusion-related energy losses required for evaluating the cooling efficiency. Adiabatic temperature change (ΔT ad ) obtained in an electric field representing long-term operation, reaches 4.91 • C for Terpo at 100 V/ μm and 3.0 • C for BSTM at 30 • C. The loss in ferroelectric hysteresis is observed to be much lower in BSTM than in PVDF Terpolymer. Hence, the cooling efficiency relative to Carnot reveals that the PVDF Terpolymer achieves a relative cooling efficiency upper bound of 4.5% at 100 V/μm, whereas BSTM ceramics reach nearly 14 times, being 62.1% at 30 V/μm. The PVDF Terpolymer outperforms BSTM ceramics in terms of adiabatic temperature change and flexibility, but not in terms of expected cooling efficiency. Considering these complementary strengths, both BSTM ceramics and PVDF terpolymers emerge as promising electrocaloric materials for advanced energy applications, including solid-state cooling and energy harvesting
Algorithme de tarification dynamique dans le secteur de la beauté
No full textInternational audienceAlgorithme de tarification dynamique dans le secteur de la beaut
A Two-Stage Cascaded Ensemble based on CRNN and Markov Chain for Sleep Apnea Detection using ECG
No full textInternational audienceSleep-related irregular breathing and apnea involve periodic and cyclical decreases or interruptions in airflow, which may occur with or without obstructions of the upper airway. With nearly a billion people affected by this sleep disorder worldwide, its screening represents a major medical issue. Its early diagnosis is all the more important as numerous studies have highlighted the correlation between the presence of an untreated sleep apnea syndrome (SAS) and neurocognitive and cardiovascular consequences. Thus, in order to speed up and improve the diagnostic management of patients, researches have been conducted towards noninvasive and portable screening methods. Some of the latter are based on the patient's cardiac activity, which is closely linked to the respiratory signal and easily recordable. In this paper, a new approach based on cascaded falseprediction-correcting ensemble using a hybrid deep model and Markov Chain is presented to detect sleep apnea events from nighttime long-term single-lead electrocardiograms (ECG), taken from the Apnea-ECG Database. The effectiveness of this approach is demonstrated through its capability to detect pathological ECG segments with a sensitivity, specificity and accuracy of 95.8%, 80% and 86%, respectively